TELETID

In my view, the first Global Dialogue on AI Governance would be successful if it achieves three mutually reinforcing outcomes, closely aligned with the mandate of the General Assembly. First, a shared and actionable understanding of priorities for global AI governance. Success would mean greater clarity on how issues such as safe, secure and trustworthy AI, human rights protection, transparency and accountability relate to capacity‑building, development needs and the Sustainable Development Goals. In particular, recognizing capacity gaps as a cross‑cutting factor across all thematic clusters would help ensure that global governance frameworks are meaningful for countries and institutions at different levels of readiness. Second, the identification of practical governance pathways. Beyond reaffirming principles, the Dialogue should surface concrete options related to standardization, interoperability and open‑source ecosystems, as well as the role of non‑centralized and distributed AI architectures. These approaches can help reduce over‑concentration, support data sovereignty, and better reflect cultural and linguistic diversity, while remaining compatible with global norms and shared standards. Third, momentum through continuity. A successful Dialogue would lay the groundwork for follow‑up mechanisms that sustain engagement beyond the inaugural session. A light platform for action or thematic follow‑up tracks—focused on areas such as AI standardization, human oversight, AI literacy for policymakers, and capacity‑building in developing countries—would help translate dialogue into ongoing cooperation and peer learning, including South‑South collaboration. Overall, success should be measured by the Dialogue's ability to connect global norms with practical, inclusive governance pathways, fostering trust, international coherence, and a development‑oriented approach to AI governance over time.

• AI capacity-building
• Interoperability of governance approaches
• Open-source software, open data and open AI models
• Transparency, accountability, and human oversight

The selected priority areas reflect a focus on practical, inclusive and development-oriented AI governance, consistent with the mandate of General Assembly resolution 79/325 and with the objective of translating global principles into actionable pathways. AI capacity-building is a central priority, as effective AI governance depends on institutional readiness, access to infrastructure, skills development and AI literacy. Capacity gaps directly affect the ability of countries-particularly developing countries-to implement safeguards related to safety, human rights, transparency and accountability. Addressing capacity-building as a cross-cutting enabler is therefore essential to closing AI divides and ensuring meaningful participation. Interoperability of governance approaches is another urgent area of engagement. Greater coherence and compatibility among national, regional and sectoral frameworks can reduce fragmentation, facilitate cooperation and support mutual trust. Interoperability, supported by shared standards, is particularly important for enabling diverse governance models to coexist while remaining aligned with international norms. Open-source software, open data and open AI models are prioritized due to their potential to promote inclusion, resilience and accessibility. Open and non-centralized AI ecosystems can help reduce over-concentration, support data sovereignty, and better reflect cultural and linguistic diversity, while also lowering barriers to entry for public institutions, educators and local innovators. Finally, transparency, accountability and human oversight are highlighted as essential foundations for trustworthy AI. These elements should be understood not only as technical requirements, but also as institutional and educational challenges that require shared standards, public-sector capacity and informed oversight. Together, these priorities emphasize a governance approach that connects global norms with practical implementation, supports sustainable development, and ensures that AI benefits are broadly shared.

While the thematic areas identified in General Assembly resolution 79/325 are comprehensive, there are a few cross-cutting and emerging issues that could be made more explicit, as they shape the effectiveness of all listed themes. First, standardization as an enabling governance mechanism deserves greater visibility as a transversal issue. While interoperability is addressed, the role of shared technical, procedural and institutional standards-particularly open and inclusive standard-setting processes-is critical to translating principles such as safety, transparency, and human oversight into practice. Without common reference points, governance risks fragmentation and uneven implementation. Second, the implications of non-centralized and distributed AI architectures represent an emerging area not fully captured by existing themes. Federated, locally governed and open-source AI systems raise important questions for governance, accountability, data sovereignty and capacity-building. At the same time, they offer opportunities to reduce over-concentration, support linguistic and cultural diversity, and enable broader participation by public institutions, educators and local innovators. Third, institutional and societal readiness, including AI literacy beyond technical communities, could be more explicitly recognized. Effective oversight, accountability and human-rights-based governance depend on informed policymakers, regulators, educators and public servants who can critically engage with AI systems. This challenge cuts across capacity-building, transparency, and trust. Finally, the environmental and resource implications of AI, including energy use, infrastructure concentration and sustainability trade-offs, are emerging issues that intersect with development priorities and equity considerations, particularly for countries with limited resources. Recognizing these cross-cutting dimensions within the Dialogue would strengthen coherence across themes, support practical implementation, and help ensure that global AI governance frameworks remain inclusive, forward-looking and responsive to diverse contexts.

Governance gaps in artificial intelligence are having both constraining and enabling effects, particularly in the education and public‑sector spheres. One of the most significant challenges relates to capacity‑building. While interest in adopting AI tools is growing, many public institutions and educational systems face limited access to infrastructure, skills, and institutional guidance. This gap affects the ability to implement safeguards related to transparency, accountability and human oversight in a meaningful way, increasing reliance on externally developed systems that are not always well adapted to local legal, cultural or linguistic contexts. A second challenge concerns interoperability and standardization. Fragmented governance approaches and the absence of shared standards make it difficult to align national initiatives with regional and global frameworks. This fragmentation creates uncertainty for public actors and educators, complicates compliance, and limits opportunities for cooperation and knowledge‑sharing across institutions and borders. At the same time, these gaps highlight important opportunities. Advances in open‑source software, open data and open AI models, as well as interest in non‑centralized and distributed AI architectures, offer promising pathways to reduce dependency on centralized solutions. These approaches can lower barriers to entry, support data sovereignty, and enable the development of AI applications that are more culturally and linguistically relevant, particularly in education and public services. Finally, there is a growing recognition that transparency, accountability and human oversight are not only technical challenges, but also institutional and educational ones. Strengthening AI literacy among policymakers, public servants and educators presents a key opportunity to improve governance outcomes over time. Overall, while governance gaps present real risks, they also create a window to shape AI governance in a way that is inclusive, interoperable and development‑oriented, if supported by sustained international cooperation and capacity‑building.

The AI Dialogue can play a catalytic role in advancing international cooperation on AI governance by serving as a trusted, inclusive and action‑oriented United Nations platform that connects diverse governance efforts and stakeholders. First, the Dialogue can help build shared understanding and coherence across fragmented initiatives. By bringing together Member States, international organizations, technical communities, academia and civil society, the Dialogue can promote convergence around common reference points—such as human‑rights‑based approaches, safety, transparency and accountability—while respecting different national contexts and levels of development. This is particularly important for enhancing interoperability and compatibility among governance approaches. Second, the Dialogue can support cooperation by linking principles to practical pathways. Through structured exchanges on standardization, capacity‑building and open infrastructures, it can help identify concrete options for implementation, especially for developing countries. Highlighting the potential of open‑source ecosystems and non‑centralized AI architectures can also foster collaboration that reduces dependency, supports data sovereignty, and enables more equitable participation in AI development and use. Third, the AI Dialogue can act as a bridge between global norm‑setting and national implementation. By sharing lessons learned, use cases and institutional experiences—particularly in education and the public sector—it can strengthen mutual learning and peer support, including South‑South and triangular cooperation. Finally, the Dialogue can advance international cooperation through continuity and follow‑up. Establishing light but sustained mechanisms, such as thematic tracks or a platform for action, would help maintain momentum between annual sessions and turn dialogue into ongoing collaboration. Overall, the AI Dialogue has the potential to strengthen international cooperation by fostering trust, reducing fragmentation, and enabling inclusive, development‑oriented AI governance that benefits all countries and communities.

The AI Dialogue can build upon a growing landscape of existing international, regional and multi‑stakeholder initiatives on AI governance, while adding distinctive value through its convening power and inclusivity. Relevant foundations include ongoing work within the United Nations system, such as UNESCO's efforts on AI ethics and capacity‑building, initiatives on digital cooperation and emerging technologies, and programmes supporting AI for sustainable development. Beyond the UN, regional and plurilateral frameworks addressing AI governance, technical standard‑setting bodies, open‑source communities, and partnerships focused on data, computing infrastructure and skills development provide important substantive inputs and practical experience. At the same time, many of these initiatives operate in parallel and fragmented ways, reflecting different mandates, levels of technical detail and regional priorities. The added value of the AI Dialogue lies precisely in its potential to connect and amplify these efforts within an inclusive global forum. First, the Dialogue can serve as a bridge for coherence and interoperability, helping relate diverse governance approaches, standards and ethical frameworks without imposing uniform solutions. This is particularly important for countries with limited capacity to engage across multiple parallel processes. Second, the Dialogue can elevate capacity‑building and development perspectives as cross‑cutting priorities, ensuring that global discussions on safety, human rights, transparency and oversight are informed by the realities of developing countries and public‑sector institutions. Third, the Dialogue can highlight and legitimize open‑source, open data and non‑centralized AI ecosystems as complementary pathways for international cooperation, innovation and inclusion, building on existing communities while situating them within global governance discussions. Finally, the AI Dialogue can add value through continuity—by fostering sustained exchanges, peer learning and follow‑up mechanisms that translate existing initiatives into longer‑term cooperation. In doing so, it can help move from fragmented efforts toward a more connected, inclusive and development‑oriented global AI governance landscape.

Different stakeholders can contribute to the AI Dialogue in distinct yet complementary ways, and the format and structure of the Dialogue should be designed to enable these contributions effectively. Member States play a central role in articulating policy priorities, sharing national experiences, and identifying governance gaps, particularly in relation to capacity‑building, human rights, and public accountability. Their engagement is essential for ensuring alignment with international law and the Sustainable Development Goals. International organizations and the United Nations system can contribute by connecting existing initiatives, offering comparative perspectives, and supporting coherence and interoperability across governance approaches. They are also well positioned to facilitate technical assistance and capacity‑building efforts. Academia and research institutions can provide independent analysis, evidence‑based insights, and evaluation of emerging issues such as standardization, non‑centralized AI architectures, and the societal and educational implications of AI. Their role is particularly important in supporting AI literacy and informed oversight. Civil society and community‑based organizations can contribute grounded perspectives on human rights impacts, inclusion, cultural and linguistic diversity, and the real‑world implications of AI systems for different communities. Technical communities and open‑source actors play a key role in sharing practical experiences on standards, interoperability, open infrastructures, and innovative governance‑by‑design approaches. To harness these contributions, the Dialogue should combine high‑level framing with structured, interactive formats. In addition to plenary discussions, short implementation‑focused sessions—such as case‑based exchanges or thematic breakouts—can help translate principles into practice. Clear thematic clustering, balanced speaker representation, and space for concrete examples would enhance inclusivity and substance. Finally, light continuity mechanisms, such as thematic follow‑up tracks or a platform for action, would allow stakeholders to remain engaged beyond the Dialogue itself, helping to transform dialogue into sustained international cooperation.

Several key voices and perspectives remain underrepresented, limiting the inclusiveness and effectiveness of existing discussions. First, actors from developing countries, particularly public‑sector institutions, educators and regulators, are often underrepresented in substantive policy shaping. While many frameworks are discussed at a global level, the realities of limited infrastructure, skills gaps and institutional capacity are not always sufficiently reflected. Their inclusion could be strengthened through targeted support for participation, regional preparatory dialogues, and capacity‑building that enables meaningful engagement rather than symbolic representation. Second, educational institutions and educators are not always fully recognized as governance stakeholders. Given their central role in AI literacy, workforce preparation and societal understanding of AI, their perspectives are critical for addressing long‑term governance challenges related to accountability, oversight and public trust. Structured engagement with universities, teacher networks and academic institutions—especially from diverse linguistic and cultural contexts—would add significant value. Third, communities working on open‑source, non‑centralized and locally governed AI systems remain underrepresented in high‑level governance discussions. These actors offer practical insights into alternative AI architectures that can support inclusion, data sovereignty and resilience, particularly for smaller institutions and countries. Creating space for their participation can enrich discussions on interoperability and standards beyond centralized models. Finally, communities affected by AI systems but not traditionally represented in technical or policy forums, including linguistic minorities and marginalized groups, are still insufficiently included. Their perspectives are essential to understanding real‑world impacts and avoiding governance approaches that inadvertently reinforce exclusion. To address these gaps, the AI Dialogue could adopt inclusive participation formats, regional and thematic outreach, and follow‑up mechanisms that sustain engagement over time. Doing so would strengthen the legitimacy, relevance and development‑oriented nature of global AI governance.

To foster meaningful and dynamic engagement, the AI Dialogue would benefit from formats that balance inclusivity, substance and practicality, while respecting time constraints and the multilateral setting. First, short implementation‑focused dialogues could complement high‑level plenaries. These could take the form of brief, moderated exchanges built around concrete cases—for example, experiences with AI standardization, non‑centralized AI architectures, or capacity‑building initiatives in education and the public sector. Structured around guiding questions, such segments can help translate principles into practice and encourage peer learning. Second, thematic multi‑stakeholder panels with clearly differentiated roles can enhance engagement. Rather than general statements, panellists could be invited to address a specific dimension of a theme—policy, technical implementation, institutional readiness, or social impact—followed by targeted reactions from other stakeholder groups. This format helps surface complementarities and tensions in a constructive way. Third, the Dialogue could experiment with interactive clustering sessions, where participants contribute short inputs aligned with predefined thematic clusters or governance challenges. Moderators could then synthesize these contributions in real time, making visible areas of convergence and divergence. This would reinforce the sense of collective ownership of outcomes. Fourth, regional or sector‑based spotlights—brief segments highlighting perspectives from developing countries, educators, public institutions, or open‑source communities—could ensure that underrepresented voices are integrated into the core discussions rather than confined to side events. Finally, to sustain engagement beyond the sessions themselves, the Dialogue could be linked to light continuity mechanisms, such as thematic follow‑up tracks, curated repositories of inputs, or voluntary communities of practice. These formats would help transform dynamic dialogue into ongoing cooperation and practical progress. Together, these engagement formats can support a Dialogue that is not only inclusive and participatory, but also forward‑looking and action‑oriented.

Standardization and interoperability initiatives represent an effective governance practice. Common technical and procedural standards help reduce fragmentation, enable cross-border cooperation, and support regulatory coherence across jurisdictions. When standards are developed through open and inclusive processes, they also allow diverse governance models to remain compatible without imposing uniform solutions. Open-source software, open data and open AI models demonstrate how governance and innovation can reinforce each other. Open AI ecosystems lower barriers to entry for public institutions, educators and smaller organizations, support adaptability to local contexts, and help reduce over-dependence on centralized actors. In particular, non-centralized and distributed AI architectures, such as locally governed or federated systems, offer promising alternatives that support data sovereignty, cultural and linguistic diversity, and institutional resilience. Capacity-building platforms and peer-learning mechanisms are critical governance enablers. Initiatives that combine technical training, policy guidance and institutional learning-especially for public servants, regulators and educators-help translate global principles into effective national and sectoral practices. Multi-stakeholder dialogue platforms that connect policymakers, academia, civil society and technical communities provide valuable spaces for exchanging lessons learned and aligning approaches. When linked to continuity mechanisms and practical outputs, such platforms can help move AI governance from fragmented efforts toward sustained international cooperation.